Algorithm Design And Optimization For Multi-echelon And Green Vehicle Routing Problems

Confronting huge online trading system as well as the great pressure of competition, logistics companies have to improve its service quality and reduce cost. Vehicle routing problem calculates reasonable distribution route for logistics companies to reduce logistics transportation cost and improve their service quality. This article focuses on multi-echelon vehicle routing problem and green vehicle routing problem.In response to multi-echelon vehicle routing problem, this paper designs a hybrid heuristic algorithm, which can solve large scale vehicle routing problem with more than two stages. In order to improve the quality of searched solution, this paper proposes a load balancing strategy, mak ing the searched solution transfer from state of unbalanced load to state of balanced load quickly. The algorithm decomposes the multi-echelon vehicle routing problem into sub-problems; In consideration of the independence of the final stage, combined with the parallel optimization, this paper designed a mixed strategy of simulated annealing algorithm to solve the vehicle routing problem in the last stage, and calculates the exact optimum solution for that in intermediate stage through linear programming. The experimental results show that, on 66 sets of benchmark data, the object value deviates from optimum solution by an average of 2.4%, and optimal solutions are found for 56.06% of the data set. Compared with the existing accurate algorithm, its final solution differs by an average of 2.8%, and this algorithm can deal with larger issues and any stage number of vehicle routing problem.For the green vehicle routing problem targeting the optimization of energy consumption, parallel annealing algorithm with hybrid strategies is proposed, considering the manipulability in reality environment, this paper chooses the vehicle load and distance as the key factors that affect vehicle energy consumption rate, and establish the relationship between the energy consumption rate and the two key factors. Based on that, we model the green vehicle routing problem. When designing the heuristic algorithm for green vehicle routing problem, the load balancing strategy which is the achievement of the multi-echelon vehicle routing problem is applied to refine the simulated annealing algorithm. We search the solution concurrently with GPU, which expanding solution searching scope effectively. The experimental results show that compared with the original algorithm, the object values are reduced by average 42.49%, and the standard deviations are reduced by average 86.41%.